Robert L. Myklebust

Modeling of the bremsstrahlung radiation produced in pure-element targets by 10-40 keV electrons

A new global relationship has been developed for predicting electron‐excited bremsstrahlung intensities over a wide range of accelerating voltages 10–40 keV, atomic numbers 4–92, and x‐ray energies 1.5–20 keV. The new relationship was determined empirically from the mathematical modeling of extensive data and is designed for calculating bremsstrahlung intensities in analytical procedures, such as those requiring peak‐to‐background measurements, where the direct measurement of the bremsstrahlung intensities is impracticable. The distribution of errors between the data and the model is symmetrical, centered around zero error with 63% of the values falling between ±10% relative error.

Determining Limits of Detection from Energy Dispersive X-ray Spectra

When comprehensive quantitative electron-excited x-ray microanalysis is performed, the analyst is often required to specify the limits of detection (appropriate to the particular analytical conditions chosen, e.g., x-ray peak, electron dose, spectrometer resolution and efficiency, etc.) for selected elemental species of interest that might be expected to be present as trace constituents.[1] If a multi-element standard is available that is similar in composition to the unknown and that contains the constituent(s) of interest at a low but measurable level (e.g., as a minor constituent where the concentration is 0.01 < C < 0.1 mass fraction), the minimum detectable concentration, CDL, can be readily estimated for the unknown. In such a dilute system, the calibration curve of k-value (where k = intensity of element “A” in the specimen/intensity of A in a pure A standard) vs. concentration is a linear function. That is, the element of interest is already so dilute that the major constituents control the matrix factors, atomic number (Z), absorption (A) and fluorescence (F), thus giving a constant value of the slope, a, in the Ziebold-Ogilvie expression relating the k-value and concentration, C:[2] (1-k)/k = a (1 – C)/C (1) The relationship between the concentration limit of detection, CDL, and the number of peak NP and background NB counts in the spectrum compared to these values for the measured minor concentration CM, is then: CDL = CM [(NP – NB)DL/(NP – NB)M] (2) It is typical practice to define CDL where the difference between the detectable peak and background is three times the standard deviation of the specimen background counts: (NP – NB)DL = 3 NB.

The R Factor: The X-Ray Loss Due to Electron Backscatter

With the evolution of electron microprobes that employ more stable electronics which permit high precision measurements of x-ray intensity, a new interest has arisen in the accuracy of methods and parameters used to compute the matrix corrections for quantitative electron probe microanalysis. One of the two components that contribute to the “atomic number correction,” part of the ZAF matrix correction procedure, is the R-factor. The R-factor is an estimate of the ratio of the total inner shell ionization which is actually produced in the specimen over that which would occur in the absence of electron backscattering. There are several formulations of the R-factor in use today. They are based on either empirical fits to experimental data on electron backscattering and backscattered electron energy distributions [1] or on data generated by a Monte Carlo simulation for electron scattering in solids [2]. In this publication we will examine several of these formulations and compare the results obtained by each method.

Quantitative Compositional Mapping with the Electron Probe Microanalyzer

Of all the techniques of electron probe microanalysis, the one that has undergone the least change over the history of the field is the technique of producing an image of the distribution of the elemental constituents of a sample, which can be termed compositional mapping. Even today with the dominance of computers for digital data collection and processing in the microprobe laboratory, most compositional mapping is still carried out with an analog procedure that is little changed from the “dot mapping” or “area scanning” technique described by Cosslett and Duncumb in 1956 [1]. The dot mapping procedure can be summarized as follows: (1) As in conventional scanning electron imaging, the beam on the cathode ray tube (CRT) is scanned in synchronism with the beam on the specimen. (2) When the beam is at a particular position on the specimen and an x-ray photon is detected with either a wavelength-dispersive (WDS) or an energy dispersive spectrometer (EDS), the corresponding beam location on the CRT is marked by adjusting the current to excite the phosphor to full brightness. (3) The white dots produced on the CRT display are continuously recorded by photographing the screen to produce the dot map.

Stereo presentation of Monte Carlo electron trajectory simulations

Electron trajectory data from Monte Carlo simulation techniques is three dimensional in nature, and thus is best represented by methods that most preserve the spatial information. Stereo plotting is a method that gives the three‐dimensional illusion effectively while not requiring any special equipment beyond what is required to make standard two‐dimensional plots. Stereo plots of electron trajectories are presented that illustrate the advantages of the spatial illusion in the context of examining in detail some of the interactions of the electron beam with planar bulk meta lic samples.

Defocus modelling correction for wavelength dispersive digital compositional mapping with the electron microprobe

A new, simplified procedure for correcting the defocusing observed in low‐magnification digital maps obtained with the electron microprobe using wavelength spectrometers is described. This procedure uses a wavelength scan of the analysed element and the geometric relationship between the specimen and the diffracting crystal to calculate a model of a standard map, which is subsequently used in the quantification of each pixel of the unknown map. The results of this new procedure are compared with the earlier method of using an experimentally obtained standard map.

The f(χ) Machine: An Experimental Bench for the Measurement of Electron Probe Parameters

In routine electron probe analysis involving sample targets that are electron opaque, flat, and conductive, the mechanisms describing the interaction of the beam electrons with the target atoms and the subsequent x-ray generation, absorption, and detection are well known. Various correction procedures are currently available for routine quantitative analysis that offer accuracies of 2 percent relative, as determined from studies of well-characterized, homogeneous standards [1].

Quantitative Electron Probe Microanalysis of Fly Ash Particles

Fly ash particles or other similar particles may be quantitatively analyzed with a flat-sample, matrix-correction method that has been modified to include the peak-to-background ratio for each element as a normalizing factor. The effects of the different matrix corrections on particles are discussed. Examples of analyses of standard reference material glass particles by both a standard matrix correction program (FRAME C) and a modified correction program (FRAME P) are presented as well as analyses of fly ash (SRM-1633).